sexp automated format & cleanup

floats
Ondřej Hruška 4 years ago
parent 3cb445a31c
commit 7272eecc48
Signed by untrusted user: MightyPork
GPG Key ID: 2C5FD5035250423D
  1. 379
      lib/spanned_sexp/src/lib.rs

@ -18,9 +18,9 @@ use std::str::{self, FromStr};
#[derive(PartialEq, Clone, PartialOrd)]
#[allow(missing_docs)]
pub enum Atom {
S(String),
I(i64),
F(f64),
S(String),
I(i64),
F(f64),
}
/// An s-expression is either an atom or a list of s-expressions. This is
@ -28,25 +28,25 @@ pub enum Atom {
#[derive(PartialEq, Clone, PartialOrd)]
#[allow(missing_docs)]
pub enum Sexp {
Atom(Atom),
List(Vec<Sexp>),
Atom(Atom),
List(Vec<Sexp>),
}
/// The representation of an s-expression parse error.
pub struct Error {
/// The error message.
pub message: &'static str,
/// The line number on which the error occurred.
pub line: usize,
/// The column number on which the error occurred.
pub column: usize,
/// The index in the given string which caused the error.
pub index: usize,
/// The error message.
pub message: &'static str,
/// The line number on which the error occurred.
pub line: usize,
/// The column number on which the error occurred.
pub column: usize,
/// The index in the given string which caused the error.
pub index: usize,
}
impl error::Error for Error {
fn description(&self) -> &str { self.message }
fn cause(&self) -> Option<&dyn error::Error> { None }
fn description(&self) -> &str { self.message }
fn cause(&self) -> Option<&dyn error::Error> { None }
}
/// Since errors are the uncommon case, they're boxed. This keeps the size of
@ -61,292 +61,291 @@ type Err = Box<Error>;
type ERes<T> = Result<T, Err>;
impl fmt::Display for Error {
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}:{}: {}", self.line, self.column, self.message)
}
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}:{}: {}", self.line, self.column, self.message)
}
}
impl fmt::Debug for Error {
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}", self)
}
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}", self)
}
}
fn get_line_and_column(s: &str, pos: usize) -> (usize, usize) {
let mut line: usize = 1;
let mut col: isize = -1;
for c in s.chars().take(pos+1) {
if c == '\n' {
line += 1;
col = -1;
} else {
col += 1;
let mut line: usize = 1;
let mut col: isize = -1;
for c in s.chars().take(pos + 1) {
if c == '\n' {
line += 1;
col = -1;
} else {
col += 1;
}
}
}
(line, cmp::max(col, 0) as usize)
(line, cmp::max(col, 0) as usize)
}
#[cold]
fn err_impl(message: &'static str, s: &str, pos: &usize) -> Err {
let (line, column) = get_line_and_column(s, *pos);
Box::new(Error {
message: message,
line: line,
column: column,
index: *pos,
})
let (line, column) = get_line_and_column(s, *pos);
Box::new(Error {
message: message,
line: line,
column: column,
index: *pos,
})
}
fn err<T>(message: &'static str, s: &str, pos: &usize) -> ERes<T> {
Err(err_impl(message, s, pos))
Err(err_impl(message, s, pos))
}
/// A helpful utility to trace the execution of a parser while testing. It will
/// be compiled out in release builds.
#[allow(unused_variables)]
fn dbg(msg: &str, pos: &usize) {
//println!("{} @ {}", msg, pos)
//println!("{} @ {}", msg, pos)
}
fn atom_of_string(s: String) -> Atom {
match FromStr::from_str(&s) {
Ok(i) => return Atom::I(i),
Err(_) => {},
};
match FromStr::from_str(&s) {
Ok(i) => return Atom::I(i),
Err(_) => {}
};
match FromStr::from_str(&s) {
Ok(f) => return Atom::F(f),
Err(_) => {},
};
match FromStr::from_str(&s) {
Ok(f) => return Atom::F(f),
Err(_) => {}
};
Atom::S(s)
Atom::S(s)
}
// returns the char it found, and the new size if you wish to consume that char
fn peek(s: &str, pos: &usize) -> ERes<(char, usize)> {
dbg("peek", pos);
if *pos == s.len() { return err("unexpected eof", s, pos) }
if s.is_char_boundary(*pos) {
let ch = s[*pos..].chars().next().unwrap();
let next = *pos + ch.len_utf8();
Ok((ch, next))
} else {
// strings must be composed of valid utf-8 chars.
unreachable!()
}
dbg("peek", pos);
if *pos == s.len() { return err("unexpected eof", s, pos); }
if s.is_char_boundary(*pos) {
let ch = s[*pos..].chars().next().unwrap();
let next = *pos + ch.len_utf8();
Ok((ch, next))
} else {
// strings must be composed of valid utf-8 chars.
unreachable!()
}
}
fn expect(s: &str, pos: &mut usize, c: char) -> ERes<()> {
dbg("expect", pos);
let (ch, next) = peek(s, pos)?;
*pos = next;
if ch == c { Ok(()) } else { err("unexpected character", s, pos) }
dbg("expect", pos);
let (ch, next) = peek(s, pos)?;
*pos = next;
if ch == c { Ok(()) } else { err("unexpected character", s, pos) }
}
fn consume_until_newline(s: &str, pos: &mut usize) -> ERes<()> {
loop {
if *pos == s.len() { return Ok(()) }
let (ch, next) = peek(s, pos)?;
*pos = next;
if ch == '\n' { return Ok(()) }
}
loop {
if *pos == s.len() { return Ok(()); }
let (ch, next) = peek(s, pos)?;
*pos = next;
if ch == '\n' { return Ok(()); }
}
}
// zero or more spaces
fn zspace(s: &str, pos: &mut usize) -> ERes<()> {
dbg("zspace", pos);
loop {
if *pos == s.len() { return Ok(()) }
let (ch, next) = peek(s, pos)?;
dbg("zspace", pos);
loop {
if *pos == s.len() { return Ok(()); }
let (ch, next) = peek(s, pos)?;
if ch == ';' { consume_until_newline(s, pos)? }
else if ch.is_whitespace() { *pos = next; }
else { return Ok(()) }
}
if ch == ';' { consume_until_newline(s, pos)? } else if ch.is_whitespace() { *pos = next; } else { return Ok(()); }
}
}
fn parse_quoted_atom(s: &str, pos: &mut usize) -> ERes<Atom> {
dbg("parse_quoted_atom", pos);
let mut cs: String = String::new();
expect(s, pos, '"')?;
loop {
let (ch, next) = peek(s, pos)?;
if ch == '"' {
*pos = next;
break;
} else if ch == '\\' {
let (postslash, nextnext) = peek(s, &next)?;
if postslash == '"' || postslash == '\\' {
cs.push(postslash);
} else {
cs.push(ch);
cs.push(postslash);
}
*pos = nextnext;
} else {
cs.push(ch);
*pos = next;
dbg("parse_quoted_atom", pos);
let mut cs: String = String::new();
expect(s, pos, '"')?;
loop {
let (ch, next) = peek(s, pos)?;
if ch == '"' {
*pos = next;
break;
} else if ch == '\\' {
let (postslash, nextnext) = peek(s, &next)?;
if postslash == '"' || postslash == '\\' {
cs.push(postslash);
} else {
cs.push(ch);
cs.push(postslash);
}
*pos = nextnext;
} else {
cs.push(ch);
*pos = next;
}
}
}
// Do not try i64 conversion, since this atom was explicitly quoted.
Ok(Atom::S(cs))
// Do not try i64 conversion, since this atom was explicitly quoted.
Ok(Atom::S(cs))
}
fn parse_unquoted_atom(s: &str, pos: &mut usize) -> ERes<Atom> {
dbg("parse_unquoted_atom", pos);
let mut cs: String = String::new();
dbg("parse_unquoted_atom", pos);
let mut cs: String = String::new();
loop {
if *pos == s.len() { break }
let (c, next) = peek(s, pos)?;
loop {
if *pos == s.len() { break; }
let (c, next) = peek(s, pos)?;
if c == ';' { consume_until_newline(s, pos)?; break }
if c.is_whitespace() || c == '(' || c == ')' { break }
cs.push(c);
*pos = next;
}
if c == ';' {
consume_until_newline(s, pos)?;
break;
}
if c.is_whitespace() || c == '(' || c == ')' { break; }
cs.push(c);
*pos = next;
}
Ok(atom_of_string(cs))
Ok(atom_of_string(cs))
}
fn parse_atom(s: &str, pos: &mut usize) -> ERes<Atom> {
dbg("parse_atom", pos);
let (ch, _) = peek(s, pos)?;
dbg("parse_atom", pos);
let (ch, _) = peek(s, pos)?;
if ch == '"' { parse_quoted_atom (s, pos) }
else { parse_unquoted_atom(s, pos) }
if ch == '"' { parse_quoted_atom(s, pos) } else { parse_unquoted_atom(s, pos) }
}
fn parse_list(s: &str, pos: &mut usize) -> ERes<Vec<Sexp>> {
dbg("parse_list", pos);
zspace(s, pos)?;
expect(s, pos, '(')?;
dbg("parse_list", pos);
zspace(s, pos)?;
expect(s, pos, '(')?;
let mut sexps: Vec<Sexp> = Vec::new();
let mut sexps: Vec<Sexp> = Vec::new();
loop {
zspace(s, pos)?;
let (c, next) = peek(s, pos)?;
if c == ')' {
*pos = next;
break;
loop {
zspace(s, pos)?;
let (c, next) = peek(s, pos)?;
if c == ')' {
*pos = next;
break;
}
sexps.push(parse_sexp(s, pos)?);
}
sexps.push(parse_sexp(s, pos)?);
}
zspace(s, pos)?;
zspace(s, pos)?;
Ok(sexps)
Ok(sexps)
}
fn parse_sexp(s: &str, pos: &mut usize) -> ERes<Sexp> {
dbg("parse_sexp", pos);
zspace(s, pos)?;
let (c, _) = peek(s, pos)?;
let r =
if c == '(' { Ok(Sexp::List(parse_list(s, pos)?)) }
else { Ok(Sexp::Atom(parse_atom(s, pos)?)) };
zspace(s, pos)?;
r
dbg("parse_sexp", pos);
zspace(s, pos)?;
let (c, _) = peek(s, pos)?;
let r =
if c == '(' { Ok(Sexp::List(parse_list(s, pos)?)) } else { Ok(Sexp::Atom(parse_atom(s, pos)?)) };
zspace(s, pos)?;
r
}
/// Constructs an atomic s-expression from a string.
pub fn atom_s(s: &str) -> Sexp {
Sexp::Atom(Atom::S(s.to_owned()))
Sexp::Atom(Atom::S(s.to_owned()))
}
/// Constructs an atomic s-expression from an int.
pub fn atom_i(i: i64) -> Sexp {
Sexp::Atom(Atom::I(i))
Sexp::Atom(Atom::I(i))
}
/// Constructs an atomic s-expression from a float.
pub fn atom_f(f: f64) -> Sexp {
Sexp::Atom(Atom::F(f))
Sexp::Atom(Atom::F(f))
}
/// Constructs a list s-expression given a slice of s-expressions.
pub fn list(xs: &[Sexp]) -> Sexp {
Sexp::List(xs.to_owned())
Sexp::List(xs.to_owned())
}
/// Reads an s-expression out of a `&str`.
#[inline(never)]
pub fn parse(s: &str) -> Result<Sexp, Box<Error>> {
let mut pos = 0;
let ret = parse_sexp(s, &mut pos)?;
if pos == s.len() { Ok(ret) } else { err("unrecognized post-s-expression data", s, &pos) }
let mut pos = 0;
let ret = parse_sexp(s, &mut pos)?;
if pos == s.len() { Ok(ret) } else { err("unrecognized post-s-expression data", s, &pos) }
}
// TODO: Pretty print in lisp convention, instead of all on the same line,
// packed as tightly as possible. It's kinda ugly.
fn is_num_string(s: &str) -> bool {
let x: Result<i64, _> = FromStr::from_str(&s);
let y: Result<f64, _> = FromStr::from_str(&s);
x.is_ok() || y.is_ok()
let x: Result<i64, _> = FromStr::from_str(&s);
let y: Result<f64, _> = FromStr::from_str(&s);
x.is_ok() || y.is_ok()
}
fn string_contains_whitespace(s: &str) -> bool {
for c in s.chars() {
if c.is_whitespace() { return true }
}
false
for c in s.chars() {
if c.is_whitespace() { return true; }
}
false
}
fn quote(s: &str) -> Cow<str> {
if !s.contains("\"")
&& !string_contains_whitespace(s)
&& !is_num_string(s) {
Cow::Borrowed(s)
} else {
let mut r: String = "\"".to_string();
r.push_str(&s.replace("\\", "\\\\").replace("\"", "\\\""));
r.push_str("\"");
Cow::Owned(r)
}
if !s.contains("\"")
&& !string_contains_whitespace(s)
&& !is_num_string(s) {
Cow::Borrowed(s)
} else {
let mut r: String = "\"".to_string();
r.push_str(&s.replace("\\", "\\\\").replace("\"", "\\\""));
r.push_str("\"");
Cow::Owned(r)
}
}
impl fmt::Display for Atom {
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
match *self {
Atom::S(ref s) => write!(f, "{}", quote(s)),
Atom::I(i) => write!(f, "{}", i),
Atom::F(d) => write!(f, "{}", d),
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
match *self {
Atom::S(ref s) => write!(f, "{}", quote(s)),
Atom::I(i) => write!(f, "{}", i),
Atom::F(d) => write!(f, "{}", d),
}
}
}
}
impl fmt::Display for Sexp {
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
match *self {
Sexp::Atom(ref a) => write!(f, "{}", a),
Sexp::List(ref xs) => {
write!(f, "(")?;
for (i, x) in xs.iter().enumerate() {
let s = if i == 0 { "" } else { " " };
write!(f, "{}{}", s, x)?;
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
match *self {
Sexp::Atom(ref a) => write!(f, "{}", a),
Sexp::List(ref xs) => {
write!(f, "(")?;
for (i, x) in xs.iter().enumerate() {
let s = if i == 0 { "" } else { " " };
write!(f, "{}{}", s, x)?;
}
write!(f, ")")
}
}
write!(f, ")")
},
}
}
}
impl fmt::Debug for Atom {
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}", self)
}
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}", self)
}
}
impl fmt::Debug for Sexp {
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}", self)
}
fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
write!(f, "{}", self)
}
}
#[cfg(test)]
@ -357,14 +356,14 @@ mod test {
fn test_hello_world() {
assert_eq!(
parse("(hello -42\n\t -4.0 \"world\") ; comment").unwrap(),
list(&[ atom_s("hello"), atom_i(-42), atom_f(-4.0), atom_s("world") ]));
list(&[atom_s("hello"), atom_i(-42), atom_f(-4.0), atom_s("world")]));
}
#[test]
fn test_escaping() {
assert_eq!(
parse("(\"\\\"\\q\" \"1234\" 1234)").unwrap(),
list(&[ atom_s("\"\\q"), atom_s("1234"), atom_i(1234) ]));
list(&[atom_s("\"\\q"), atom_s("1234"), atom_i(1234)]));
}
#[test]
@ -388,7 +387,7 @@ mod test {
#[test]
fn test_space_in_atom() {
let sexp = list(&[ atom_s("hello world")]);
let sexp = list(&[atom_s("hello world")]);
let sexp_as_string = sexp.to_string();
assert_eq!("(\"hello world\")", sexp_as_string);
assert_eq!(sexp, parse(&sexp_as_string).unwrap());
@ -418,6 +417,6 @@ mod test {
fn sexp_size() {
// I just want to see when this changes, in the diff.
use std::mem;
assert_eq!(mem::size_of::<Sexp>(), mem::size_of::<isize>()*5);
assert_eq!(mem::size_of::<Sexp>(), mem::size_of::<isize>() * 5);
}
}

Loading…
Cancel
Save